Hey everyone,
Just wanted to let you know about our software called Abmarl which has recently been published to the Journal of Open Source Software. We built Abmarl to work with RLlib for training MARL. We developed many features from personal experience in MARL and with RLlib. We don’t replace RLlib (how could we, it’s our main dependency); instead, we try to make it easier to setup MARL workflows that utilize RLlib for training.
You can find our more in our docs and check out the code from our repo.
Hopefully it helps you as much as it has helped our team! We are actively developing and welcome constructive feedback, so please check us out.
Thanks!
